I. Group I Introns
RNA molecules with catalytic activity are called ribozymes or RNA enzymes (Cech, T. R., Ann. Rev. Biochem. 59:543-568 (1990). The Tetrahymena thermophila precursor rRNA contains an intron (a ribozyme) capable of catalyzing its own excision. This ribozyme is one of a class of structurally related Group I introns.
The splicing activity of the modified T. thermophila intron requires the presence of a guanosine cofactor and a divalent cation, either Mg.sup.++ or Mn.sup.++, and occurs via two sequential transesterification reactions (FIG. 1). First, a free guanosine is bound to the ribozyme and its 3' hydroxyl group is positioned to attack the phosphorus atom at the 5' splice site. The guanosine is covalently attached to the intron sequence and the 5' exon is released. Second, the phosphodiester bond located at the 3' splice site undergoes attack from the newly freed 3' hydroxyl group of the 5' exon, resulting in production of the ligated exon sequences. The excised intron subsequently undergoes a series of transesterification reactions, involving its 3' hydroxyl group and internal sequences, resulting in the formation of shortened circular forms.
These successive reactions are chemically similar and appear to occur at a single active site. The reactions of self-splicing are characterized by the formation of alternative RNA structures as differing RNA chains are each brought to form similar conformations around the highly conserved intron. Splicing requires the alignment of the intron-exon junctions across a complementary sequence termed the "internal guide sequence" or IGS.
The first cleavage at the 5' splice site requires the formation of a base-paired helix (P1) between the IGS and sequences adjacent the splice site. The presence of a U:G "wobble" base-pair within this helix defines the phosphodiester bond that will be broken in the catalytic reaction of the ribozyme. After cleavage of this bond, a portion the P1 helix is displaced and a new helix, P10, is formed due to complementarity between the IGS and sequences adjacent the 3' splice site. An invariant guanosine residue precedes the phosphodiester at the 3' splice site, similar to the portion of the P1 sequence that it is displacing. Thus, ligation of the exons occurs in a reverse of the first cleavage reaction but where new exon sequences have been substituted for those of the intron. It may be noted that intron circularization reactions subsequent to exon ligation also involve base-pairing of 5' sequences across the IGS, and attack mediated by the 3' hydroxyl group of the intron's terminal guanine residue (Been, M. D. et al., "Selection Of Circularization Sites In A Group I IVS RNA Requires Multiple Alignments Of An Internal Template-Like Sequence," Cell 50:951 (1987)).
II. Catalytic Activities
In order to better define the structural and catalytic properties of the Group I introns, exon sequences have been stripped from the "core" of the T. thermophila intron. Cech, T. R. et al., WO 88/04300, describes at least three catalytic activities possessed by the Tetrahymena intron ribozyme: (1) a dephosphorylating activity, capable of removing the 3' terminal phosphate of RNA in a sequence-specific manner, (2) an RNA polymerase activity (nucleotidyl transferase), capable of catalyzing the conversion of oligoribonucleotides to polyribonucleotides, and (3) a sequence-specific endoribonuclease activity.
Isolated ribozyme activities can interact with substrate RNAs in trans, and these interactions characterized. For example, when truncated forms of the intron are incubated with sequences corresponding to the 5' splice junction, the site undergoes guanosine-dependent cleavage in mimicry of the first step in splicing. The substrate and endoribonucleolytic intron RNAs base-pair to form helix P1, and cleavage occurs after a U:G base-pair at the 4th-6th position. Phylogenetic comparisons and mutational analyses indicate that the nature of the sequences immediately adjacent the conserved uracil residue at the 5' splice site are unimportant for catalysis, provided the base-pairing of helix P1 is maintained (Doudna, J. A. et al., Proc. Natl. Acad. Sci. USA 86: 7402-7406 (1989)).
The sequence requirements for 3' splice-site selection appear to lie mainly within the structure of the intron itself, including helix P9.0 and the following guanosine residue which delineates the 3' intron boundary. However, flanking sequences within the 3' exon are required for the formation of helix P10 and efficient splicing, as shown by mutational analysis (Suh, E. R. et al., Mol. Cell. Biol. 10:2960-2965 (1990)). In addition, oligonucleotides have been ligated in trans, using a truncated form of the intron, and "external" guide sequence and oligonucleotides which had been extended by a 5' guanosine residue. The substrate oligonucleotides corresponding to 3' exon sequences were aligned solely by the formation of P10-like helices on an external template, prior to ligation (Doudna, J. A. et al., Nature 339:519-522 (1989)).
The cleavage activity of ribozymes has been targeted to specific RNAs by engineering a discrete "hybridization" region into the ribozyme, such hybridization region being capable of specifically hybridizing with the desired RNA. For example, Gerlach, W. L. et al., EP 321,201, constructed a ribozyme containing a sequence complementary to a target RNA. Increasing the length of this complementary sequence increased the affinity of this sequence for the target. However, the hybridizing and cleavage regions of this ribozyme were integral parts of each other. Upon hybridizing to the target RNA through the complementary regions, the catalytic region of the ribozyme cleaved the target. It was suggested that the ribozyme would be useful for the inactivation or cleavage of target RNA in vivo, such as for the treatment of human diseases characterized by the production of a foreign host's RNA. However, ribozyme-directed trans-splicing, (as opposed to trans-cleavage) was not described or suggested.
The endoribonuclease activities (the cleavage activities) of various naturally-occurring ribozymes have been extensively studied. Analysis of the structure and sequence of these ribozymes has indicated that certain nucleotides around the cleavage site are highly conserved but flanking sequences are not so conserved. This information has lead to the design of novel endoribonuclease activities not found in nature. For example, Cech and others have constructed novel ribozymes with altered substrate sequence specificity (Cech, T. R. et al., WO 88/04300; Koizumi, M. et al., FEBS Lett. 228:228-230 (1988); Koizumi, M. et al., FEBS Lett. 239:285-288 (1988); Haseloff, J. et al., Nature 334:585-591 (1987); and Heus, H. A. et al., Nucl. Acids Res. 18:1103-1108 (1990)). From early studies of the self-cleaving plant viroids and satellite RNAs (Buzayan, J. M. et al., Proc. Natl. Acad. Sci. USA 83:8859-8862 (1986), guidelines for the design of ribozymes that are capable of cleaving other RNA molecules in trans in a highly sequence specific have been developed (Haseloff, J. et al., Nature 334:585-591 (1988)). However, these constructs were unable to catalyze efficient, targeted trans-splicing reactions.
The joining of exons contained on separate RNAs, that is, trans-splicing, occurs in nature for both snRNP-mediated and self-catalyzed group I and group II introns. In trypanosome and Caenorhabditis elegans mRNAs, common 5' leader sequences are transcribed from separate genes and spliced to the 3' portions of the mRNAs (Agabian, N., Cell 61:1157-1160 (1990); Hirsh, D. et al., Mol. Biol. Rep. 14:115 (1990). These small "spliced leader" RNAs (slRNAs) consist of the 5' exon fused to sequences that can functionally substitute for U1 snRNA in mammalian snRNP-splicing extracts.
Also, both the group I and group II self-splicing introns are capable of exon ligation in trans in artificial systems (Been, M. D. et al., Cell 47:207-216 (1986); Galloway-Salvo, J. L. et al., J. Mol. Biol. 211:537-549 (1990); Jacquier, A. et al., Science 234:1099-1194 (1986); and Jarrell, K. A. et al., Mol. Cell Biol. 8:2361-2366 (1988)). Trans-splicing occurs in vivo for group II introns in split genes of chloroplasts (Kohchi, T. et al., Nucl. Acids Res. 16:10025-10036 (1988)), and has been shown for a group I intron in an artificially split gene in Escherichia coli (Galloway-Salvo, J. L. et al., J. Mol. Biol. 211:537-549 (1990)). In the latter case, a bacteriophage T4 thymidylate synthase gene (td) containing a group I intron was divided at the loop connecting the intron helix P6a. Transcripts of the td gene segments were shown to undergo trans-splicing in vitro, and to rescue dysfunctional E. coli host cells. Known base-pairings (P3, P6 and P6a) and possible tertiary interactions between the intron segments, allowed correct assembly and processing of the gene halves.
In vitro, the Tetrahymena ribozyme is capable of catalyzing the trans-splicing of single-stranded model oligoribonucleotide substrates. Four components were necessary: ribozyme, 3' single-stranded RNA, 5' exon and GTP. A shortened form of the Tetrahymena ribozyme (L-21 ScaI IVS RNA), starting at the internal guide sequence and terminating at U.sub.409 has been used in such a reaction (Flanegan, J. B. et al., J. Cell. Biochem. (Supp.)12 part D:28 (1988)). Attack by GTP at the 5' splice site released the 5' exon which was then ligated by the ribozyme to the 3' exon in a transesterification reaction at the 3' splice site.
The in vivo use of ribozymes as an alternative to the use of antisense RNA for the targeting and destruction of specific RNAs has been proposed (Gerlach, W. L. et al., EP321,201; Cotten, M., Trends Biotechnol. 8:174-178 (1990); Cotten, M. et al., EMBO J. 8:3861-3866 (1989); Sarver, N. et al., Science 247:1222-1225 (1990)). For example, expression of a ribozyme with catalytic endonucleolytic activity towards an RNA expressed during HIV-1 infection has been suggested as a potential therapy against human immunodeficiency virus type 1 (HIV-1) infection (Sarver, N. et al., Science 247:1222-1225 (1990); Cooper, M., CDC AIDS Weekly, Apr. 3, 1989, page 2; Rossi, J. J., Abstract of Grant No. 1RO1AI29329 in Dialog's Federal Research in Progress File 265). However, such attempts have not yet been successful.
In a study designed to investigate the potential use of ribozymes as therapeutic agents in the treatment of human immunodeficiency virus type 1 (HIV-1) infection, ribozymes of the hammerhead motif (Hutchins, C. J. et al., Nucl. Acids Res. 14:3627 (1986); Keese, P. et al., in Viroids and Viroid-Like Pathogens, J. S. Semancik, ed., CRC Press, Boca Raton, Fla., 1987, pp. 1-47) were targeted to the HIV-1 gag transcripts. Expression of the gag-targeted ribozyme in human cell cultures resulted in a decrease (but not a complete disappearance of) the level of HIV-1 gag RNA and in antigen p24 levels (Sarver, N. et al., Science 247:1222-1225 (1990)). Thus, the medical effectiveness of Sarver's ribozyme was limited by its low efficiency since any of the pathogen's RNA that escapes remains a problem for the host.
Another problem with in vivo ribozyme applications is that a high ribozyme to substrate ratio is required for ribozyme inhibitory function in nuclear extracts and it has been difficult to achieve such ratios. Cotton et al. achieved a high ribozyme to substrate ration by microinjection of an expression cassette containing a ribozyme-producing gene operably linked to a strong tRNA promoter (a polymerase III promoter) in frog oocytes, together with substrate RNA that contains the cleavage sequence for the ribozyme (Cotton, M. et al., EMBO J. 8:3861-3866 (1989). However, microinjection is not an appropriate method of delivery in multicellular organisms.
The in vivo activity of ribozymes designed against mRNA coding for Escherichia coli .beta.-galactosidase has been reported (Chuat, J. C. et al., Biochem. Biophys. Res. Commun. 162:1025-1029 (1989)). However, this activity was only observed when the ribozyme and target were transfected into bacterial cells on the same molecule. Ribozyme activity was inefficient when targeted against an mRNA transcribed from a bacterial F episome that possessed the target part of the .beta.-galactosidase gene.
Thus, current technological applications of ribozyme activities are limited to those which propose to utilize a ribozyme's cleavage activity to destroy the activity of a target RNA. Unfortunately, such applications often require complete destruction of all target RNA molecules, and/or relatively high ribozyme:substrate ratios to ensure effectiveness and this has been difficult to achieve. Most importantly, the modified ribozymes of the art are not capable of efficient, directed trans-splicing.
Accordingly, a need exists for the development of highly efficient ribozymes and ribozyme expression systems. Especially, the art does not describe an effective means in which to destroy an existing RNA sequence or to alter the coding sequence of an existing RNA by the trans-splicing of a new RNA sequence into a host's RNA.